Noise generated due to flow over corrugated pipes has extensively been studied, because of its presence in various engineering applications. Flexible risers used in the oil and gas industry are one of these applications, where noise is produced due to corrugations present at the innermost layer of the riser. Dry-gas flow through a corrugated pipe generates an unstable shear layer over each of the corrugations. Under certain conditions, this shear layer can roll up into discrete vortices, which impinge on the downstream cavity edge, producing pressure pulsations. When the frequency of impingement matches with one of the natural pipe frequencies, a ‘lock-in’ mechanism causes intensification of the associated noise and vibrations, which can have serious implications on the structural integrity of the corrugated pipe. This phenomenon is called Flow Induced Pulsations (FIP). Liquid injection in a corrugated pipe with gas flow has shown the potential to mitigate whistling caused by FIP. A number of mechanisms play a role in the whistling mitigation. It is shown in literature that the filling of cavities due to liquid film or rivulets present on the pipe wall is an important mechanism. The cavity filling results in an altered cavity geometry, which changes the shear layer dynamics. The effect of liquid and gas flow rate on the cavity filling, however, remains unknown. The present work aims to study the liquid filling behaviour of a single cavity, due to a gas flow driven liquid film.
An experimental setup is constructed which consists of a rectangular channel containing a cavity. Liquid is injected in the channel as a film at the channel wall, which is driven upward by gas flow. The liquid film thickness upstream of the cavity is measured first, followed by the cavity filling itself. Both measurements are performed using a Laser Induced Fluorescence (LIF) based technique. After several post-processing steps, the film thickness and cavity filling are quantified.
At low liquid flow rates, a partial film is created at the channel wall with a dry patch at the center. Increasing the liquid flow rate results in a full film covering the entire channel width. The film thickness varies in the transverse direction due to the presence of localized horseshoe shaped disturbance waves on the liquid film. Overall, the film cross-sectional area increases with increasing liquid flow rate, either due to an increase in film thickness of full films, or an increase in width of the partial films. The liquid film fills up the cavity, accumulating mainly at the upstream edge. The downstream edge remains relatively free from liquid. The amount of filling is found to increase with increasing local film thickness at the measurement position away from the channel side-wall. Changes in the cavity geometry due to liquid filling are estimated based on the cavity effective depth and length. The effective cavity depth does not significantly change and only decreases by 8% with a 25% increase in filling ratio. However, the effective length decreases substantially by 30% with a 25% increase in filling ratio. This could lead to a mode parameter value (ratio of cavity length to incoming gas momentum thickness), such that it does not fall in the range where whistling is observed, resulting in mitigation of whistling.

A liquid film falling vertically along a wall results in the formation of waves at the liquid-gas interface. The principal forces in these flows are gravity, viscosity, and surface tension, which are characterized by the Reynolds number, Re and the Kapitza number, Ka. Experimental characterization of the film flows for high viscosity liquids (Ka in the order of 1) at low Re » in the order of 10 was performed for flow along the inner circumference of a vertical pipe. The three-dimensional liquid-gas interfacial profile is quantitatively reconstructed using the laser induced fluorescence technique. The waves observed had higher surface steepness at the front when compared with the back of the wave crest. The film thickness at the back of the wave crest is higher than at the front, resembling the streak-like waves observed in the literature for low-viscosity liquids (Ka in the order of 1000) at comparable Re. When the lateral surfaces of the waves are in contact, merging in the transverse direction is observed. The experimentally determined mean film thickness was approximately 15 % lower than the Nusselt’s flat film thickness, while the experimental values of average wave velocity was approximately 4 times higher than the Nusselt’s velocity for the corresponding film thickness. On the time-averaged film thickness field, a transverse variation in the film thickness, called as ridges and valleys, were observed. These transverse variations are found to be unsteady in their behaviour. Probability density distributions of the film thickness measurements showed two distinctive time-dependant patterns for the distribution of film thickness values in the ridges.